Dynamic Property Bus for Autoscaling Cloud Instances
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In a distributed cloud environment with autoscaling, it is challenging to propagate application properties to an unknown number of instances, leading to difficulties in updating all running instances efficiently and avoiding outages.
Innovation Solution
A method is introduced that generates a dynamic property bus interface to write and propagate application property values to a database, which are then read and updated by each instance's scheduler, ensuring real-time updates across all instances without requiring instance recycling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional endpoint-based property propagation is used, then properties can be pushed to known instances, but it fails to reach all instances in an autoscaling environment where the number of instances is unknown
Solution Approach 1:
Instead of pushing properties from the central service to instances (traditional approach), the patent inverts the approach by having instances pull properties from the database. Each instance executes a scheduler that periodically reads the latest property values directly from the database, ensuring all instances receive updates regardless of when they were created or scaled.
Solution Approach 2:
Each application instance autonomously retrieves its own property values by executing a scheduler that reads from the database. This self-service mechanism eliminates the need for the central service to track and push properties to each instance individually, making the system adaptable to dynamic scaling while ensuring complete property propagation.
2Reliability
If application properties are updated by recycling instances, then all instances are updated, but this causes service outages and interruptions
Solution Approach 1:
The patent implements a scheduler in each instance that proactively and periodically reads property values from the database before any updates are needed. This preliminary action ensures that when property changes occur in the database, all running instances will detect and apply these changes during their next scheduled read, without requiring instance recycling or causing service interruptions.
3Reliability
If manual property propagation to each instance is implemented, then all instances can be updated, but this requires significant developer intervention and is inefficient
Solution Approach 1:
Each application instance autonomously retrieves its own property values by executing a scheduler that reads from the database. This self-service mechanism eliminates the need for the central service to track and push properties to each instance individually, making the system adaptable to dynamic scaling while ensuring complete property propagation.
Solution Approach 2:
The scheduler in each instance provides feedback by periodically checking the database for property changes and automatically applying updates when detected. This automated feedback loop eliminates manual developer intervention while ensuring all instances receive complete property propagation, as each instance independently monitors and applies updates.
Data Source
AI summary
In one example, a method for updating an application that has multiple instances running in a distributed cloud environment is disclosed. The method may include the operation of generating a dynamic property bus interface. The method may write an application property value through the dynamic property bus interface to a database for storage and propagate the property value to the multiple instances of the application in the distributed cloud environment.